Huntington's Disease (HD): Neurodegeneration of Brodmann's Primary Visual Area 17 (BA17)

Abstract

Huntington's disease (HD), an autosomal dominantly inherited polyglutamine or CAG repeat disease along with somatomotor, oculomotor, psychiatric and cognitive symptoms, presents clinically with impairments of elementary and complex visual functions as well as altered visual‐evoked potentials (VEPs). Previous volumetric and pathoanatomical post‐mortem investigations pointed to an involvement of Brodmann's primary visual area 17 (BA17) in HD. Because the involvement of BA17 could be interpreted as an early onset brain neurodegeneration, we further characterized this potential primary cortical site of HD‐related neurodegeneration neuropathologically and performed an unbiased estimation of the absolute nerve cell number in thick gallocyanin‐stained frontoparallel tissue sections through the striate area of seven control individuals and seven HD patients using Cavalieri's principle for volume and the optical disector for nerve and glial cell density estimations. This investigation showed a reduction of the estimated absolute nerve cell number of BA17 in the HD patients (71 044 037 ± 12 740 515 nerve cells) of 32% in comparison with the control individuals (104 075 067 ± 9 424 491 nerve cells) (Mann–Whitney U‐test; P < 0.001). Additional pathoanatomical studies showed that nerve cell loss was most prominent in the outer pyramidal layer III, the inner granular layers IVa and IVc as well as in the multiform layer VI of BA17 of the HD patients. Our neuropathological results in BA17 confirm and extend previous post‐mortem, biochemical and in vivo neuroradiological HD findings and offer suitable explanations for the elementary and complex visual dysfunctions, as well as for the altered VEP observed in HD patients.

Introduction

Huntington's disease (HD), an autosomal dominantly inherited, progressive and currently not causally treatable polyglutamine or CAG repeat disease 9, 54, 55, 56, 63, 66, 71, 72, 74, 75, presents clinically with somatomotor, oculomotor and psychiatric symptoms, cognitive decline, and in its advanced phases also with a severe weight loss 6, 20, 21, 22, 45, 54, 55, 56, 63, 66, 74, 75. Neuropsychological studies, in addition, have demonstrated a wide range of symptoms pointing to an affection of the central visual system in HD patients: dysfunctions of the perception and encoding of simple forms and faces 15, 26, 64, 65; impairments of visuospatial 5, 27, 32, 39, 40, 51, visuomnestic 39, 40 and visuodiscriminative functions 39; impaired visual attention 68 and visuomotor coordination 46 as well as naming deficits for visual objects 24, 50. Neurophysiological investigation of the visual pathway by means of visual‐evoked potentials (VEP), in addition, disclosed reduced amplitudes of their early and late wave components 10, 23, 33, 47. At the moment, the morphological counterparts of these neuropsychological and neurophysiological HD abnormalities are widely unknown.

The mutated HD gene is located on chromosome 4p16.3 and harbors meiotically unstable CAG repeats that encode the protein huntingtin with its abnormal polyglutamine expansion 9, 45, 63, 66, 73, 74, 75. Although sequences of 6–34 CAG repeats are not sufficient to cause the clinical phenotype of HD, CAG triplet repeats of 28 and more, however, behave instable during meiosis and are prone to conversion into symptomatic HD mutations. Sequences longer than 34 CAG repeats constitute pathological expansions causing clinical HD symptoms, whereby 35–40 CAG repeats are associated with an incomplete penetrance and 41 or more CAG repeats provoke the fully developed clinical picture of HD 9, 45, 63, 66, 74, 75.

The macroscopic aspect of the brain of patients in the advanced stages of HD is characterized by a severe atrophy of the neostriatum (ie, caudate nucleus and putamen), reduction of the cerebral white matter, enlargement of the lateral ventricles as well as a considerable atrophy of the frontal, parietal, temporal and occipital cerebral lobes 3, 6, 11, 16, 19, 20, 21, 22, 36, 37, 38, 54, 55, 56, 63, 71, 72, 73. Along with a severe neuronal loss in the neostriatum, marked neurodegeneration is also commonly present in select nuclei of the thalamus, in the cerebellum, brainstem as well as in specific neuronal layers of circumscribed areas of the neo‐ and allocortex of HD patients 3, 6, 11, 16, 19, 20, 21, 22, 36, 37, 38, 54, 55, 56, 63, 71, 72, 73.

Previous post‐mortem volumetric studies and pathoanatomical investigations already pointed to an involvement of Brodmann's primary visual area 17 (BA17) in the occipital lobe of HD patients 22, 38. HD is frequently regarded as a disease whereby the caudate nucleus is the primarily affected target with neuronal loss that gradually encompasses the whole neostriatum with subsequent diffuse spreading through anterograde and/or retrograde degenerative mechanisms. The anterograde or retrograde brain expansion of pathological processes requires the existence of intact anatomical tracts between vulnerable brain regions. Experimental investigations in nonhuman primates demonstrated a widespread longitudinal termination of corticostriatal efferent projections originating from the outer and inner pyramidal layers III and V of the frontal, temporal, parietal and occipital lobes to the neostriatum in contrast and to the best of our knowledge no corticostriatal projections originating from BA17 have been demonstrated so far 2, 30, 35, 57, 79, 80. Therefore, the changes involving BA17 in HD could be interpreted as an early onset and primary cortical neuronal degeneration rather than a secondary process due to the initial neurodegeneration of the neostriatum. To further characterize this potential primary cortical center of the degenerative process in HD, we estimated the absolute number of neurons present in BA17 of HD patients using an unbiased stereological approach and applied Cavalieri's principle for volume and the optical disector (OD) for nerve and glial cell (GC) density estimations 7, 8, 17, 18, 20, 21, 22, 46, 48, 61, 76, 77, 78.

Patients and Methods

Patients

For the stereological investigation of this study, we evaluated the left BA17 of seven patients in the advanced clinical stages of HD (three women, four men; age at HD onset: 40.57 ± 15.37 years: duration of HD: 11.86 ± 4.53 years; mean age at death: 52.43 ± 13.69 years) (Table 1) and of seven age‐ and gender‐matched controls (three women, four men; mean age at death: 54.43 ± 12.71 years) without any recorded neuropsychiatric diseases (Table 2) 20, 21, 22. For the purposes of identification of layer‐specific pathoanatomy, we investigated BA17 of these controls and four additional individuals without neuropsychiatric diseases (two women, two men; mean age at death: 56.57 ± 6.56 years). Owing to the severe neuronal loss in the neostriatum and circumscribed areas of the cerebral cortex of our HD patients, this study could not be performed “blinded” to the clinical or neuropathological diagnosis (Figures 1 and 2) 22.

Table 1.

Overview of the Huntington's disease (HD) patients investigated by means of stereological methods. Abbreviations: BA17 = Brodmann's primary visual area 17; CE(G) = predicted coefficient of error for the estimation of the individual absolute glial cell number; CE(N) = predicted coefficient of error for the estimation of the individual absolute nerve cell number; CV = coefficient of variation; F = female; G = estimated individual total glial cell number in the primary visual area 17; GD = estimated glia cell density: 1/mm³; GI = glial index; HD = Huntington's disease; M = male; m = arithmetic mean; N = estimated individual total nerve cell number; ND = estimated nerve cell density: 1/mm³; SD = standard deviation; U = unidentified cells; VC = corrected estimated volume: mm³; VU = uncorrected estimated volume: mm³

Case	8	9	10	11	12	13	14	m	sd	cv
Diagnosis	HD	HD	HD	HD	HD	HD	HD
Gender	F	M	M	F	M	F	M
Age at death	36	40	41	54	62	62	72	52.43	13.69	0.26
Age at HD onset	24	29	29	36	56	45	65	40.57	15.37	0.38
Duration of HD	12	11	12	18	6	17	7	11.86	4.53	0.38
N	73 484 853	80 447 286	55 611 592	86 384 125	69 734 048	79 048 840	52 597 517	71 044 037	12 740 515	0.18
CE (N)	3.11	2.74	3.19	2.87	2.97	3.29	2.96	3.02	0.19	0.06
G	79 477 380	75 138 047	63 382 069	62 857 766	71 967 736	70 618 099	61 031 326	69 210 348	6 968 441	0.1
CE (G)	3.74	3.54	3.73	4.20	3,65	3,71	3.44	3.72	0.24	0.06
GI	1.08	0.93	1.14	0.73	1.03	0.89	1.16	1.00	0.15	0.15
U	6 897 056	8 258 974	5 784 352	8 976 538	7 682 254	8 157 220	6 155 299	7 415 956	1 175 239	0.16
VU—mm3	2 607	2 898	2 471	2 787	2 913	2 799	2 320	2 685	226	0.08
VC—mm3	4 495	5 914	3 861	5 465	n.d.	n.d.	4 639	4 875	815	0.17
ND—N/mm3	28 186	27 760	22 505	30 994	23 940	28 248	22 761	26 342	3 266	0.12
GD—G/mm3	30 485	25 928	25 650	22 553	24 707	25 235	26 313	25 839	2 391	0.09

Number of HD patient, clinical diagnosis, gender (F—female; M—male), age at death (years), age at HD onset (years), duration of HD (years), estimated absolute neuron number N in BA17, coefficient of error CE(N) for the estimation of the absolute neuron number in BA17, estimated absolute glial cell number G in BA17, coefficient of error CE(G) for the estimation of the absolute glial cell number in BA17, glial index GI, unidentified cells U in BA17, estimated uncorrected volume of BA17 (VU), estimated corrected volume of BA17 (VC), estimated neuronal density ND in BA17, estimated glial cell density GD in BA17. Arithmetic means (m), standard deviation (sd) and coefficients of variation (cv) are provided in bold and italicized numbers.

Table 2.

Overview of the control individuals investigated by means of stereological methods. Abbreviations: BA17 = Brodmann's primary visual area 17; CE(G) = predicted coefficient of error for the estimation of the individual absolute glial cell number; CE(N) = predicted coefficient of error for the estimation of the individual absolute nerve cell number; CV = coefficient of variation; F = female; G = estimated individual total glial cell number in the primary visual area 17; GD = estimated glia cell density: 1/mm³; GI = glial index; M = male; m = arithmetic mean; N = estimated individual total nerve cell number; ND = estimated nerve cell density: 1/mm³; SD = standard deviation; U = unidentified cells; VC = corrected estimated volume: mm³; VU = uncorrected estimated volume: mm³

Case	1	2	3	4	5	6	7	m	sd	cv
Diagnosis	Control	Control	Control	Control	Control	Control	Control
Gender	M	F	M	F	M	F	M
Age at death	36	46	49	53	58	64	75	54.43	12.71	0.23
N	105 274 792	119 504 718	110 726 089	104 824 214	92 464 858	93 581 796	102 149 002	104 075 067	9 424 491	0.09
CE (N)	2.66	2.52	2.83	2.76	2.67	2.94	2.41	2.68	0.18	0.07
G	90 195 688	95 911 889	103 172 337	85 643 312	85 242 165	106 734 042	88 155 167	93 579 229	8 601 089	0.09
CE (G)	3.60	3.52	3.67	3.25	3.48	3.44	3.35	3.47	0.14	0.04
GI	0.86	0.80	0.93	0.82	0.92	1.14	0.86	0.90	0.11	0.12
U	7 329 904	9 356 704	7 644 682	8 188 229	7 824 893	8 178 392	8 329 024	8 121 690	647 873	0.08
VU—mm3	3 340	3 340	3 552	3 220	3 590	3 611	3 126	3 397	191	0.06
VC—mm3	6 302	4 985	6 701	n.d.	6 528	n.d.	n.d.	6 129	780	0.13
ND—N/mm3	31 521	35 782	31 177	32 356	25 753	25 917	32 677	30 740	3 667	0.12
GD—G/mm3	27 006	28 718	29 050	26 599	23 742	29 560	28 201	27 553	1 991	0.07

Number of control individual, clinical diagnosis, gender (F = female; M = male), age at death (years), estimated absolute neuron number N in BA17, coefficient of error CE(N) for the estimation of the absolute neuron number in BA17, estimated absolute glial cell number G in BA17, coefficient of error CE(G) for the estimation of the absolute glial cell number in BA17, glial index GI, unidentified cells U in BA17, estimated uncorrected volume of BA17 (VU), estimated corrected volume of BA17 (VC), estimated neuronal density ND in BA17, estimated glial cell density GD in BA17. Arithmetic means (m), standard deviation (sd) and coefficients of variation (cv) are provided in bold and italicized numbers.

Figure 1.

Neuronal loss in Brodmann's primary visual area 17 (BA17) in Huntington's disease (HD). A. Thick gallocyanin‐stained frontal tissue section through the caudal pole of the occipital lobe of a 50‐year‐old male control individual without neuropsychiatric diseases in his medical records showing Brodmann's primary visual area 17 (BA17) and its borders with the adjacent Brodmann's parastriate area 18 (broken lines). Note the prominent pale, cell‐poor and myelin‐rich Gennari stripe in the inner granular layer IVb of BA17 (asterisks) (4, 13, 70, 80). Arrowheads point to the basophilic multiform layer VI. B. BA17 of a 36‐year‐old female HD patient (HD case 8; Table 2). The striking pallor of and neuronal loss in (1) the outer pyramidal layer III and (2) inner granular layer IVc, as well as (3) the reduction of the multiform layer VI are detectable even at this low magnification. Broken lines mark the sharp boundaries with Brodmann's parastriate area 18 and asterisks Gennari stripe in the inner granular layer IVb of BA17. Bars in A and B indicate 1200 μm. (A, B: Nissl staining with gallocyanin).

Figure 2.

Predominant neuronal loss in layers III, IVa, IVc and VI in Brodmann's primary visual area 17 (BA17) in Huntington's disease (HD).

A. Magnified section of a thick gallocyanin‐stained frontal tissue section through Brodmann's primary visual area 17 (BA17) of a 65‐year‐old female control individual. The “calcarine” cortex type of BA17 shows a strong reduction of the outer and inner pyramidal layers III and V, well‐developed granular layers II and IV with a very high neuronal density, a strongly basophilic multiform layer VI and a subdivision of the inner granular layer into layers IVa–c. Note the Gennari stripe in the inner granular layer IVb (asterisks) 4, 13, 70, 80. B. BA17 of a 36‐year‐old female HD patient (HD case 8; Table 2): (1) Severe neuronal loss and pallor of the outer pyramidal layer III of BA17. (2) Considerably reduced nerve cell density and pallor of the inner granular layer IVa of BA17. (3) Reduced nerve cell density and pallor of the inner granular layer IVc of BA17. (4) Narrowed multiform layer VI of BA17 owing to nerve cell loss. C. Picture detail of the outer pyramidal layer III of BA17 of the 65‐year‐old female control individual. D. Picture detail of the outer pyramidal layer III of BA17 of the 36‐year‐old female HD patient (HD case 8; Table 2) showing pallor and reduced neuronal density of this BA17 layer. E. Picture detail of the multiform layer VI of BA17 of the 65‐year‐old female control individual. F. Picture detail of the multiform layer VI of BA17 of the 36‐year‐old female HD patient (HD case 8; Table 2): narrowing of and neuronal loss in this BA17 layer. Asterisks indicate Gennari stripe. Bars in A and B indicate 150 μm, while bars in C–F indicate 75 μm (A–F: Nissl staining with gallocyanin).

Tissue preparation and staining

Autopsies were performed in accordance with the laws of the Federal Republic of Germany and the evaluations of the brains were approved by the Ethical Board of the Faculty of Medicine at the University of Würzburg.

Subsequent to their immersion fixation in 10% formalin for at least 3 months, the 18 left cerebral hemispheres were embedded in gelatin and then completely cut into 600–700 μm thick frontoparallel serial sections. In each instance every third section from the frontoparallel tissue sections underwent Nissl staining with gallocyanin and was used for pathoanatomical and quantitative stereological investigations of BA17 17, 18, 19, 20, 21, 22. Because of the staining, dehydration and mounting procedures, the frontoparallel serial tissue sections underwent shrinkage of approximately 20% 17, 18, 22.

The outlines, borders and cytoarchitectonic features of BA17 were identified in these Nissl‐stained serial sections by means of a wide‐field stereomicroscope at 7.5× magnification and according to acknowledged anatomical criteria (Figures 1 and 2) 4, 13, 70, 80. Estimation of the volume, nerve cell (NC) and GC numbers in BA17 was performed by UR (Tables 1 and 2) and investigation of its laminar pathoanatomy by UR and HH (Figures 1 and 2).

Volume estimations according to the Cavalieri principle

The volume of BA17 of the control individuals and HD patients was estimated in accordance to the Cavalieri principle (Tables 1 and 2) 7, 8, 20, 21, 22, 48, 49, 61, 76, 77, 78. To this end area estimation of BA17 was performed on every third gallocyanin‐stained frontoparallel serial section (D = 3) through the occipital calcarine sulcus along its rostrocaudal axis up to the occipital pole by means of the point counting method and a transparent square lattice test system that was placed on the frontoparallel sections and fixed by an adhesive tape. This test system consisted of a grid of 50 × 50 squares each of which was equivalent to a point counting area A = 4 mm². All crossing points (ΣP) of this grid that hit BA17 were counted for the estimation of the total area of BA17, and the thickness (T) of each frontoparallel section was determined with an Olympus BH microscope 18, 20, 21, 22. Following Cavalieri's principle, we calculated the uncorrected total volume of BA17 (VU) (Tables 1 and 2) 7, 8, 17, 18, 20, 21, 22, 48, 49, 61, 76, 77, 78.

The estimated volume of BA17 was corrected where possible in the control individuals and HD patients by an individual cortex‐specific shrinkage factor (SF) 22. For this purpose, we estimated the volumes of the cerebral cortex in the unstained and stained left cerebral hemispheres and applied our transparent test system with its grid of 50 × 50 squares (point counting area of each square A = 4 mm²) on every third (D = 3) unstained and gallocyanin‐stained, shrunken frontoparallel tissue section through the left cerebral hemispheres 17, 18, 21, 22. Subsequent to counting all points of the test system that crossed the cerebral cortex (ΣP), we calculated the volumes of the cerebral cortex in the unstained and gallocyanin‐stained left cerebral hemispheres. Division of the volume of the cerebral cortex estimated in gallocyanin‐stained frontoparallel tissue sections by the volume of the cerebral cortex estimated in unstained frontoparallel tissue sections revealed the SF 20, 21, 22 and correction of VU by the SF yielded the corrected estimated volume of BA17 (VC) (Tables 1 and 2) 22.

Estimations of absolute NC and GC numbers by means of OD

The gallocyanin‐stained frontoparallel sections through BA17 were fixed on the mechanical stage of an Olympus BH microscope (oil immersion objective 40/1.0 combined with 10 × 10 wide‐field eyepieces; final magnification: 400×). Starting at a random point at the left upper border of BA17 and in accordance with the systematic random sampling scheme, we meandered at fixed intervals through BA17 of the control individuals and HD patients and analyzed all visual fields of interest covering the striate area by means of differently spaced ODs. Twenty‐five squares (equivalent to a visual field area VA of the OD of 0.00391 mm²) of the 5 × 5 mm ocular grid subdivided by 10 × 10 lines inserted into a wide‐field (10×) eyepiece of this microscope were used as unbiased counting frame for counting of NCs and 16 squares (equivalent to a visual field area VA of the OD of 0.0025 mm²) as unbiased counting frame for the counting of GCs. In accordance with the principles of the OD, all profiles of NCs or GCs appearing in their unbiased counting frames were recorded during focusing within a fixed depth or height of the OD (H) of 0.0297 mm 7, 8, 17, 20, 21, 22, 48, 49, 61, 76, 77, 78.

Because the profiles of small granule and astroglial cells in BA17 could not always be differentiated with certainty in gallocyanin‐stained tissue sections, these indistinguishable cells were recorded as unidentified cells U (Tables 1 and 2) 22. Likewise, because the profiles of GCs could not be unequivocally differentiated into astroglial, oligodendroglial or microglial cells GCs in BA17 were counted in total (Tables 1 and 2) 20, 21, 22.

Estimation of NC densities in BA17 (ND) was based in the control individuals on the analysis of the visual field areas of 385 ± 45 OD and counting of 1404 ± 193 NC and in the HD patients on the investigation of the visual field areas of 366 ± 56 OD and recording of 1109 ± 140 NC. Estimation of GC densities in BA17 (GD) was based in the control individuals on the analysis of 409 ± 50 OD and counting of 833 ± 70 GCs and in the HD patients on the assessment of 385 ± 46 OD and recording of 732 ± 87 GC (all values expressed as arithmetic means ± standard deviation).

Subsequent to the assessment procedures, we calculated the estimated density of neurons (ND) and the estimated GC density (GD) in BA17 of the control individuals and HD patients (Tables 1 and 2) 7, 8, 20, 21, 22, 48, 49, 61, 76, 77, 78. The product of VU and ND revealed the estimated absolute number of NCs (N) in BA17 and the product of VU and GD the estimated absolute number of GCs (G) in BA17 (Tables 1 and 2) 7, 8, 20, 21, 22, 48, 49, 61, 76, 77, 78.

As a reliable predictor of the precision of the estimated absolute NC number in BA17, we calculated an individual coefficient of error CE(N) and as a predictor of the precision of the estimated absolute GC number in BA17 an individual coefficient of error CE(G) (Tables 1 and 2) 20, 60, 61, 62. Finally, division of G by N revealed the glial index (GI) (Tables 1 and 2) 20, 21, 22, 36.

Statistical analysis

Differences in the estimated morphometrical parameters of BA17 between the control individuals and HD patients were assessed by means of the nonparametric Mann–Whitney U‐test (BiAS for Windows version 9.14, Epsilon, Darmstadt, Germany). Kendall's nonparametric rank correlation coefficient tau (τ) was applied to examine bivariate linear correlations between the age of the control individuals and HD patients, on the one hand, and the estimated morphological parameters, on the other hand (BiAS for Windows version 9.14, Epsilon). Kendall's nonparametric rank correlation coefficient tau was also used to correlate the estimated absolute neuronal number in BA17 of our HD patients with the onset at disease onset and duration of HD, both of which could be regarded as indirect estimators of the clinical severity of HD.

Results

BA17 in gallocyanin‐stained thick tissue sections

BA17 (or area OC according to von Economo and Koskinas) settles the depth and the banks of the calcarine sulcus on the medial surface of the occipital lobe as well as its most caudal portions (Figures 1 and 2). Our gallocyanin‐stained frontoparallel tissue sections through the calcarine sulcus along its rostrocaudal axis up to the occipital pole (i) enabled the unequivocal delineation of BA17 and the identification of its layers and sharp boundaries with Brodmann's parastriate area 18; (ii) revealed the six‐layered calcarine cortex type of BA17 that displays the most differentiated laminar structure of all neocortical areas of the human brain; (iii) showed a strong reduction of the outer and inner pyramidal layers III and V and well‐developed granular layers II and IV similar to other sensory human cortical areas; (iv) displayed the highest neuronal density in the outer and inner granule layers II and IV and a strong basophilia of the multiform layer VI; and (v) allowed the subdivision of the inner granular layer into layers Iva–c. These cytoarchitectonic features as well as the presence and abrupt endings of the prominent pale, cell‐poor and myelin‐rich Gennari stripe in its layer IVb that corresponds to the outer band of Baillarger allow (i) the reliable localization of BA17 and (ii) its accurate delineation from Brodmann's visual parastriate area 18 that forms a belt around BA17 (Figures 1 and 2) 4, 13, 70, 80.

Volume, NC and GC numbers in BA17 in HD

Our quantitative, stereological investigations revealed:

• (i) An estimated uncorrected volume VU of BA17 of 2.685 > 2685 mm³ in the HD patients and of 3.397 > 3397 mm³ in the control individuals (Mann–Whitney U‐test; P < 0.001) (Tables 1 and 2).
• (ii) An estimated corrected volume VC of BA17 of 4.875 > 4875 mm³ in the HD patients and of 6.129 > 6129 mm³ in the control individuals (Mann–Whitney U‐test; P < 0.07) (Tables 1 and 2).
• (iii) An estimated nerve cell density ND of 26.342 > 26 342 and 3.266 > 3266 NCs/mm³ in BA17 of the HD patients and 30.740 > 30 740 and 3.667 > 3667 NCs/mm³ in BA17 of the control individuals (Mann–Whitney U‐test; P < 0.04) (Tables 1 and 2).
• (iv) An estimated absolute NC number N of 71 044 037 ± 12 740 515 NCs in BA17 of the HD patients and 104 075 067 ± 9 424 491 NCs in BA17 of the control individuals (Mann–Whitney U‐test; P < 0.001) (Figure 3; Tables 1 and 2).
• (v) An estimated GC density GD of 25 839 ± 2 391 GCs/mm³ in BA17 of the HD patients and 27 553 ± 1991 GCs/mm³ in the control individuals (Mann–Whitney U‐test; P < 0.13) (Tables 1 and 2).
• (vi) An estimated absolute GC number G of 69 210 347 ± 6 968 441 GCs in BA17 of the HD patients and 93 579 229 ± 8 601 089 GCs in the control individuals (Mann–Whitney U‐test; P < 0.001) (Tables 1 and 2).
• (vii) An estimated GI of 1.00 ± 0.15 in BA17 of the HD patients and 0.90 ± 0.11 in BA17 of the control individuals (Mann–Whitney U‐test; P < 0.21) (Tables 1 and 2).

Figure 3.

The estimated absolute nerve cell number in Brodmann's primary visual area 17 (BA17) in Huntington's disease (HD). Estimated absolute nerve cell number in BA17 of age‐ and gender‐matched control individuals (filled circles) and Huntington's disease (HD) patients (filled squares). Horizontal lines represent arithmetic means. Estimation of the absolute neuronal numbers is based on the application of Cavalieri's principle for volume and the optical disector for nerve density estimations 7, 8, 18, 20, 21, 22, 48, 49, 61, 76, 77, 78 and revealed a reduction by 32% in the HD patients (71 044 037 ± 12 740 515 nerve cells) as compared with the control individuals (104 075 067 ± 9 424 491 nerve cells) (*** Mann–Whitney U‐test; P < 0.001) (Tables 1 and 2). In all seven HD patients studied BA17 nerve cell loss was concentrated in the outer pyramidal cell layer III, the inner granulary layers IVa and IVc, as well as in the multiform layer VI (Figures 1, 2).

Calculation of Kendall's rank correlation coefficient tau (τ) revealed no statistically significant linear correlations between the age of the control individuals and HD patients, on the one hand, and the estimated morphological parameters of BA17, on the other hand. In addition, the estimated absolute neuronal number in BA17 of our HD patients did not correlate significantly with the age at onset of HD and duration of HD.

Pathoanatomical findings in BA17 in HD

In the seven HD patients studied the occipital lobe was atrophic, the calcarine sulcus widened and the cortical ribbon of the striate area reduced. Although a diffuse NC loss in the apparently intact layers of BA17 of our HD patients could not be excluded, the brunt of NC loss involved the outer pyramidal cell layer III, the inner granular layers IVa and IVc as well as the multiform layer VI throughout their entire extent (Figures 1 and 2).

Owing to a severe neuronal loss BA17 outer pyramidal layer III of our HD patients was pale throughout its extent and showed a sharp border to the well‐preserved outer granular layer II, which normally is indistinct (Figures 1 and 2). Because they exhibited considerably reduced NC densities, the inner granular layers IVa and IVc likewise were pale when compared with our control individuals. BA17 multiform layer VI of our HD patients was consistently smaller than that of the control individuals studied (Figures 1 and 2). This BA17 layer VI neuronal loss brought the cone‐shaped NC columns in our HD patients to light, which normally are masked by surrounding nerves and therefore cannot be recognized with certainty in control individuals (Figures 1 and 2).

Discussion

Neurodegeneration of BA17 in HD

In this study we present for the first time unbiased stereological findings in BA17 of patients who died after protracted HD. These findings confirm and extend previous volumetric and pathoanatomical investigations that pointed to an involvement of BA17 in the degenerative process of HD 22, 38 and are consistent with post‐mortem biochemical (ie, elevated lactate) and in vivo neuroradiological (ie, atrophy) changes observed in BA17 of HD patients 28, 46, 47, 53, 64. Our re‐investigation of BA17 was performed on unconventional thick gallocyanin‐stained tissue sections through the primary visual striate area and is based on the systematic application of Cavalieri's principle for volume and the OD for NC and GC density estimation 7, 8, 17, 20, 21, 22, 48, 49, 61, 76, 77, 78. Unbiased estimation of absolute NC numbers revealed a statistically significant reduction by 32% in BA17 of the HD patients (71 044 037 ± 12 740 515 NCs) in comparison with the control individuals (104 075 067 ± 9 424 491 NCs). This degree of NC loss in the occipital striate area is in excellent agreement with the overall neuronal loss of 33% previously estimated by us in the cerebral cortex of five patients of the same HD sample studied here (ie, patients 8–11, 14; Table 2) and underlines the reliability of our quantitative investigations 19, 22. As in a previous study 22, pathoanatomical investigations of BA17 showed that NC loss in the HD patients occurred predominantly in its outer pyramidal cell layer III, inner granular layers IVa and IVc, and in the multiform layer VI. Based on the application of Cavalieri's principle, our volume estimations disclosed a reduction of the uncorrected and corrected volumes of BA17 of the HD patients by 21% in comparison with the control individuals studied. This estimated volume reduction of 21% conforms to the volume loss of 17% in BA17 assessed in HD brains of the C. and O. Vogt brain collection (Düsseldorf, Germany) by Lange and Aulich 38. The present estimation of neuron loss shows that the volume loss or atrophy of the striate cortex in HD patients is not alone due to shrinkage of the neuropil and/or the NCs but reflects to a definite, intrinsic premature death of neurons.

Possible pathological mechanisms of layer‐specific neurodegeneration of BA17 and other areas of the cerebral neo‐ and allocortex in HD

The neostriatum has been regarded for more than hundred years as the main and initial target of the neurodegenerative process of HD 3, 11, 16, 18, 20, 21, 22, 36, 37, 38, 54, 55, 56, 71, 72, 73. Nearly 30 years of neostriatal NC loss is known to follow consistent topographical and chronological orders, which represent the base of the widely acknowledged neuropathological Vonsattel grading system of the HD‐related brain pathology 56, 71, 72, 73. However, increasing evidence indicates that the cerebral cortex also undergoes significant degeneration in HD, which might actively participate in the widespread dispersion of the degenerative process causing symptoms characteristically occurring in HD (ie, psychiatric symptoms, cognitive decline) 3, 6, 11, 16, 20, 21, 22, 36, 38, 53, 54, 55, 56, 71, 72, 73. Well documented are macroscopical atrophy encompassing the four cerebral lobes with regional differentiated thinning of the cortex and atopistic predominant laminar neuronal loss 3, 6, 11, 16, 19, 20, 22, 38, 54, 71, 72. So far layer‐specific or predominant laminar cortical neuronal loss has been demonstrated (i) in the frontal lobe (outer pyramidal layer III, inner pyramidal layer V, multiform layer VI) 16, 19, 20, 22, 53, 71, 72; (ii) in the primary somatosensory cortex (outer pyramidal layer III, inner granular layer IV and multiform layer VI) 19, 22; (iii) in the temporal and parietal association cortices (outer pyramidal layer III) 19, 22; (iv) in the allocortical transentorhinal and entorhinal regions (deep layer pri‐γ) 3, 71, 72; and (v) is also suggested by findings in BA17 (outer pyramidal layer III, inner granular layers IVa and IVc, and multiform layer VI) of HD patients 22, present study). That the cerebral cortex undergoes degenerative changes in HD is well established; however, questions pertaining to the pathogenetic processes remain unanswered:

• (i) When, during the gradual striatal worsening of the neurodegenerative process that characteristically occurs in HD, is the neo‐ or allocortex involved, and whether either of this involvement is primary or secondary?
• (ii) Is there a correlation between the layer‐specific or predominant laminar neuronal loss in distinct areas of the neo‐ and allocortex of HD patients and the laminar neuronal population projecting to or receiving afferents from definite subcortical centers (ie, neostriatum, thalamus)?
• (iii) Are there primary centers of degeneration located within the cerebral cortex or some of its areas from which the pathological process propagates through transneuronal, anterogradely or retrogradely spreading degenerative mechanisms?

Answers to these unresolved questions will (i) lead to improved insights into the origin and pathological mechanisms of the degenerative process of HD; (ii) help to decipher the essential principles and mechanisms of its topographical and chronological spread throughout the brain; and (iii) also offer conclusive explanations of the enigmatic phenomenon of the selective vulnerability of distinct cortical and subcortical brain sites for the neurodegenerative process of HD.

Taking into account the layer‐specific or predominant laminar affection of the associative prefrontal cortex 16, 19, 20, 22, 53, 71, 72, the allocortical transentorhinal and entorhinal regions 3, 71, 72, and BA17 22 present study) we will now reconsider the possible scenarios of the early events of the neurodegenerative process of HD and the possible pathoanatomical mechanisms leading to the involvement of the cerebral cortex.

The prefrontal cortex outer pyramidal layer III, inner pyramidal layer V and multiform layer VI have been repeatedly demonstrated to undergo layer‐specific neuronal loss in HD 19, 20, 22, 53, 71, 72. The prefrontal cortex (i) is reciprocally interconnected via efferent and afferent pathways with the thalamic mediodorsal nucleus 14, 29, 42, 80, a subcortical target of the HD‐related neuronal loss 19, 20. Cortical neurons of the outer pyramidal layer III send projections to associated cortical areas and to the neostriatum 2, 80, whereas the inner granular layer IV receives thalamocortical projections from the mediodorsal nucleus 14, 80. Neurons from layers V and VI of the prefrontal cortex 14, 80, which undergo neurodegeneration in HD 16, 19, 20, 22, 53, 71, 72, 73, send projections back to the thalamic mediodorsal nucleus. In view of these corticosubcortical interconnectivities, damage to the outer pyramidal layer III in HD could be (i) the long‐distance effect of a retrogradely spreading neuronal loss occurring early in associated areas of the cerebral cortex and/or the neostriatum or (ii) represent a primary cortical degenerative event that secondarily triggers degeneration in the associated cortical areas and in the neostriatum via an anterograde expansion. Although the apparent intactness of the prefrontal inner granular layer IV in HD strongly points to well‐preserved afferents originating from the mediodorsal thalamic nucleus, damage to the prefrontal inner pyramidal layer V and multiform layer VI in HD could reflect (i) an early cortical degeneration that takes place independently from striatal degeneration and induces thalamic pathological changes via corticothalamic projections or (ii) the pathological outcome of the retrogradely transmitted disease process emanating from the affected thalamic mediodorsal nucleus.

The NCs in the deep pri‐γ layer of the allocortical transentorhinal and entorhinal regions are especially vulnerable and might undergo subtotal loss in HD 3. They have no direct efferent or afferent connections with the severely affected neostriatum but receive substantial afferent input from the prefrontal cortex and the dopaminergic ventral tegmental area of the midbrain 1, 2, 25, 43, 52. Because these layer‐specific prefrontal projections can serve as trails enabling the anterograde and/or retrograde transneuronal spread of degenerative mechanisms, transentorhinal and entorhinal pri‐γ degeneration may be an additional primary pathological focus that subsequently stimulates neurodegeneration in the prefrontal cortex or may be the consequence of an anterograde spreading prefrontal damage in HD. Owing to their projections to the transentorhinal and entorhinal pri‐γ layer, the dopaminergic nuclei of the midbrain ventral tegmental area may also be involved in the degenerative process of HD and therefore should be re‐investigated in HD.

The neostriatum of nonhuman primates receives widespread cortical projections originating predominantly from the inner pyramidal layer V of the frontal, temporal, parietal and occipital cortices. Notably, BA17 is not interconnected with the neostriatum 2, 30, 35, 57, 79. The absence of direct connections between BA17 and the neostriatum implies that neuronal degeneration in BA17 of HD patients is most likely independent from that involving the neostriatum. Neurodegeneration in HD in particular affects the outer pyramidal layer III, the inner granulary layers IVa and IVc as well as the multiform layer VI of BA17. The NCs of the outer pyramidal layer III of BA17 receive retinal input via the visual pathway and its final relais in the thalamic lateral geniculate body 31, 41, 42 and they project to ipsilateral temporal and/or occipital visual cortices 41, 42 and to the contralateral BA17 13. The NCs of the inner granular layers layer IVa and IVc of BA17, likewise, receive visual pathway information via the optic radiation and the six‐layered lateral geniculate body; are reciprocally connected with BA17 of the contralateral hemisphere; and also emit efferents to further visual cortices 13, 80). The NCs of the damaged BA17 layer VI are reciprocally interconnected with the lateral geniculate body 13, 31, 41, 80 and the claustrum 13, 31, 41, 80, send callosal fibers to the contralateral BA17 80 or receive afferent input from other extrastriate visual cortical areas 13 and from the visual subnuclei of the pulvinar 13. Following our pathoanatomical concepts, the existence of these cortical and subcortical connections of the predominantly affected BA17 layers III, IVa, IVc and VI suggests that our hypothesized extrastriatal primary focus of degeneration not only operates in BA17 but also has adverse effects for cortical and subcortical regions interconnected with these neuronal layers of BA17. Our pathoanatomical findings that all six layers of the lateral geniculate body in all of our HD patients apparently were we‐preserved (UR, personal communication) suggests that this important visual relais station of the thalamus might not be among the degenerative process of HD.

In accordance with the current neuropathological knowledge and pathoanatomical interpretations, the layer‐specific or predominant laminar neuronal loss in distinct areas of the cerebral cortex of HD patients could reflect (i) a remote secondary effect, which is induced by sick NCs in the initially affected striatal and extrastriatal brain regions and anterogradely or retrogradely conveyed to interconnected cortical areas via fiber tracts or (ii) a primary pathological event in the pathological process of HD that has subordinated detrimental effects for anatomically interconnected subcortical regions. The second pathoanatomical interpretation and explanation of the layer‐specific or predominant laminar neuronal loss in distinct areas of the cerebral cortex as a primary event is compatible with the previously proposed hypothesis that HD represents a multifocal or polytopic human neurodegenerative disease 22, 38, 71.

The possible clinical relevance of neurodegeneration of BA17 in HD

Incoming visual information flow from the retina is conveyed to the six‐layered thalamic lateral geniculate body and from there via the optic radiations to the external and internal granular layers II and IV, and external pyramidal layer III of BA17 in the occipital lobe and ultimately reaches a mosaic of anatomically and functionally related subcortical relais stations as well as to extrastriate visual areas in the temporal, parietal and occipital lobes 12, 13, 34, 41, 42, 69. Therefore, BA17 represents the major port of entry from the human central visual system; is functionally mainly concerned with the recognition, analysis and processing of the basic features of objects of the visual world and together with select visual fields in the parietal cortex and the lateral geniculate body and pulvinar; and plays an important role in the attention networks of the human brain 34, 41, 42, 44, 69. The intimately interconnected cortical Brodmann areas 18 and 19 subserve recognition of the patterns, colors and motions of the objects of the visual world and their naming 12, 34, the downstream located visually related cortices of the temporal lobe perception and recognition, visual memory and analysis of motion of objects of the visual world 12, 34, 42, 58, 59, 67, while the visual cortices of the parietal lobe allow identification and analysis of the spatial relationships among objects of the visual world and are additionally crucial for visuomotor control 34, 42, 59, 69. In view of the current knowledge in the field of the functional neuroanatomy of the human central visual system we conclude (i) that the basic [ie, dysfunctions of perception and encoding of simple forms and faces 15, 26, 50, 51; deficits of visual attention 68; visuodiscriminative deficits 39] and complex [ie, visuospatial dysfunctions 5, 27, 32, 39, 40, 51; impairment of visuomnestic functions 39, 40; naming deficits for objects of the visual world 24, 50] visual dysfunctions of HD patients cannot be straightforwardly and unilaterally attributed to the affection of the neostriatum as has been practiced for nearly hundred years when it came to explain other nonmotor HD disease symptoms 54, 56; (ii) that the visual symptoms occurring in HD are due to distinct, multifocal involvement of this hierarchically organized sensory system of the brain; and (iii) that damage to BA17 contributes substantially to the pathogenesis of the basic and complex visual dysfunctions in HD.

Normal latencies and reduced amplitudes of the early components of the VEPs have been recognized already in the pre‐genetic area of HD and are regarded as the typical electrophysiological VEP alterations in HD patients since more than 30 years 10, 23, 33, 47. Because the generators of these early VEP components have been located in the occipital calcarine sulcus covered by BA17 67, the pathological changes in the striate area demonstrated in this and other HD studies 28, 38, 46, 47, 53 most likely play a causal pathophysiological role in the evolution of VEP alterations in patients suffering from HD 10, 23, 33, 47.